Instead of one-off prompts, feed your AI a persistent knowledge base with company data, sales playbooks, and territory info. This "intelligence layer" provides crucial context, enabling the AI to perform complex, tailored sales tasks effectively and consistently.
Don't just use AI for one-way output. Close the loop by regularly feeding it data on what worked—booked meetings, positive replies, effective messaging. This creates a flywheel where the AI's intelligence layer gets progressively smarter, tightening processes and improving future prospecting results.
Chats in LLMs are temporary. To give your AI a permanent memory, store key instructions, playbooks, and processes as markdown documents within the AI's project files. This creates a stable intelligence layer that the AI always references, and the format is portable enough to be moved to other LLMs.
When feeding data to an AI for prospecting, don't limit it to your assigned accounts. The speaker exports all 19,000 company accounts. This allows the AI to surface high-signal accounts that might be unassigned or were mistakenly overlooked during territory allocation, ensuring no opportunity is missed.
When exporting CRM data for AI analysis, include all deals, not just your current pipeline. This allows the AI to incorporate data from deals that were closed-lost months ago, identifying accounts that may be ready for a renewed conversation and preventing it from mistakenly suggesting active deals.
Instead of explaining sales methodologies from scratch, the speaker copies the full transcripts from his company's internal training courses directly into the AI. The AI ingests this proprietary knowledge, creating a playbook it can apply to specific prospecting tasks like problem hypothesizing and messaging.
The speaker has evolved his AI usage. He relies on the AI for the heavy lifting of data analysis, signal identification, and problem hypothesizing. However, the final, nuanced act of crafting the outreach email is a human task, leveraging the seller's own expertise for maximum impact.
When an AI suggests a bad-fit account, don't just discard it. The speaker dictates the reason for the error back to the AI and explicitly instructs it to update its permanent "outbound process" document. This trains the AI's core logic to avoid similar mistakes in the future, creating a smarter system over time.
While an LLM is great for sifting data, it's a poor system of record because it's slow and difficult to query during a live cold call. The speaker uses AI to generate insights and account notes but immediately transfers that data into the CRM (HubSpot) to ensure it's instantly accessible when needed.
The most critical step is the last one. After a prospecting session, the speaker instructs the AI to review their entire chat, identify any mistakes or inefficiencies, and then rewrite its own core process documents. This self-correction ensures the system becomes smarter and more efficient with every use.